Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I am Partha Ghosh, an AI/ML Engineer with 3+ years building end-to-end recommendation systems and scalable backend architectures. I design AI-driven decision-support systems, develop cloud-based data pipelines, and deploy production ML models using Python, PyTorch, and TensorFlow. I excel at architecting secure, scalable systems that process complex datasets and deliver real-time recommendations through robust APIs. In my recent roles, I collaborated with researchers, product teams, and domain experts to translate complex requirements into technical solutions, delivering scalable pipelines and high-uptime services while contributing to peer-reviewed publications and competitive problem-solving outcomes.…I am Partha Ghosh, an AI/ML Engineer with 3+ years building end-to-end recommendation systems and scalable backend architectures. I design AI-driven decision-support systems, develop cloud-based data pipelines, and deploy production ML models using Python, PyTorch, and TensorFlow. I excel at architecting secure, scalable systems that process complex datasets and deliver real-time recommendations through robust APIs. In my recent roles, I collaborated with researchers, product teams, and domain experts to translate complex requirements into technical solutions, delivering scalable pipelines and high-uptime services while contributing to peer-reviewed publications and competitive problem-solving outcomes.

Partha Ghosh

AI Engineer, AI Developer, Data Scientist, +3





I am Partha Ghosh, an AI/ML Engineer with 3+ years building end-to-end recommendation systems and scalable backend architectures. I design AI-driven decision-support systems, develop cloud-based data pipelines, and deploy production ML models using Python, PyTorch, and TensorFlow. I excel at architecting secure, scalable systems that process complex datasets and deliver real-time recommendations through robust APIs. In my recent roles, I collaborated with researchers, product teams, and domain experts to translate complex requirements into technical solutions, delivering scalable pipelines and high-uptime services while contributing to peer-reviewed publications and competitive problem-solving outcomes.…I am Partha Ghosh, an AI/ML Engineer with 3+ years building end-to-end recommendation systems and scalable backend architectures. I design AI-driven decision-support systems, develop cloud-based data pipelines, and deploy production ML models using Python, PyTorch, and TensorFlow. I excel at architecting secure, scalable systems that process complex datasets and deliver real-time recommendations through robust APIs. In my recent roles, I collaborated with researchers, product teams, and domain experts to translate complex requirements into technical solutions, delivering scalable pipelines and high-uptime services while contributing to peer-reviewed publications and competitive problem-solving outcomes.

Available to hire

I am Partha Ghosh, an AI/ML Engineer with 3+ years building end-to-end recommendation systems and scalable backend architectures. I design AI-driven decision-support systems, develop cloud-based data pipelines, and deploy production ML models using Python, PyTorch, and TensorFlow. I excel at architecting secure, scalable systems that process complex datasets and deliver real-time recommendations through robust APIs.

In my recent roles, I collaborated with researchers, product teams, and domain experts to translate complex requirements into technical solutions, delivering scalable pipelines and high-uptime services while contributing to peer-reviewed publications and competitive problem-solving outcomes.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

English

Fluent

German

Intermediate

Hindi

Advanced

Bengali

Fluent

Work Experience

Machine Learning Engineer at German Cancer Research Center (DKFZ)

April 1, 2024 - Present

Architected and deployed a full-stack AI-powered platform for the Self-Supervised Learning for 3D Medical Imaging Challenge; designed scalable cloud backend infrastructure on HPC clusters; developed robust APIs and data processing pipelines enabling thousands of model submissions with 99.9% uptime; implemented data validation and preprocessing workflows for large-scale datasets (>4.8 million volumes); built automated ML pipelines integrating diverse data sources and model architectures; collaborated with domain experts to translate complex requirements into technical solutions, contributing to peer-reviewed publications; achieved 3rd place in FOMO25 Challenge.

Machine Learning Engineer at Aurivus

March 1, 2023 - February 1, 2024

Designed and implemented AI-driven recommendation engines for automated component identification and geometric property prediction, improving prediction accuracy by 35%. Developed end-to-end ML pipelines for large-scale 3D data processing (AWS), processing datasets exceeding 100GB. Built automated workflows and APIs for real-time data processing and seamless integration with downstream applications. Engineered production-ready ML systems with a focus on scalability, maintainability, and security, reducing processing time by 60%.

Research Assistant at Autonomous Vision Group, University of Tübingen

October 1, 2021 - September 30, 2022

Designed and developed an AI-powered semantic recommendation system for scientific papers, integrating NLP models with backend infrastructure. Built and optimized ML models using Python, PyTorch, and HuggingFace Transformers for multi-class classification; improved model accuracy from 58% to 69% through systematic experimentation. Implemented data preprocessing pipelines and validation for text data, and explored model compression (knowledge distillation) to reduce inference size by 75% while maintaining 95% of baseline performance.

Research Assistant at Autonomous Vision Group

October 1, 2021 - September 1, 2022

Designed and developed an AI-powered semantic recommendation system for scientific papers, integrating NLP models (BERT variants) with backend infrastructure. Built and optimized machine learning models using Python, PyTorch, and HuggingFace Transformers for multi-class classification, improving model accuracy from 58% to 69% through systematic experimentation and hyperparameter optimization. Implemented data preprocessing pipelines and validation frameworks for text data, ensuring robust model training and evaluation. Explored model compression techniques (knowledge distillation) to optimize inference performance for resource-constrained environments, reducing model size by 75% while maintaining 95% of baseline performance. Worked closely with cross-functional research team to iterate on system architecture and incorporate user feedback into technical solutions.